Brand love: conceptual and empirical investigation of a holistic causal model

Explicating and specifying the origins of brand love, as well as how it affects consumer behavior, establishes vital insights into how brand managers might reap favorable economic consequences from promoting brand love effectively. Therefore, this article presents and validates a holistic, causal model of brand love that accounts for brand stimulus features and the internal, mental processes of consumers, along with the behavioral outcomes of their resulting brand love. Using both qualitative and quantitative approaches, the authors propose and test seven antecedents (including three mediators) and four consequences: Functional and sensory brand uniqueness emerge as indirect antecedents of brand love; brand satisfaction, brand fit with the inner self, and personal experiences are direct antecedents. Contrary to expectations, communicative uniqueness and brand pleasure are not influential factors. This study also verifies four desirable behavioral consequences of brand love: brand loyalty, willingness to pay a price premium, word-of-mouth intentions, and forgiveness of brand mistakes. These findings offer several theoretical and managerial implications.


Introduction
Consumers frequently report intense emotional relationships with brands, comparable to feelings of interpersonal love Fournier 1998). Such deep consumer-brand relationships influence consumer behavior in various ways and even can reach extremes, as when Harley-Davidson fans express their lifelong devotion to the brand by getting the company's logo tattooed on their bodies. Empirical studies thus show that brand love results in positive outcomes, such as brand loyalty and willingness to pay a price premium (Albert and Merunka 2013;Rossiter 2012). Considering the ways in which these outcomes promise to enhance the performance and economic value of companies (Rossiter 2012), it is unsurprising that the brand love phenomenon attracts substantial interest from practitioners (Roberts 2004) and appears in a growing stream of research (Albert and Merunka 2013;Heinrich et al. 2012;Sajtos et al. 2020;Tsai 2013). Yet several important questions remained unaddressed. In particular, brand love studies tend to feature partial models and focus on isolated effects, without establishing a comprehensive, integrative model of the multiple antecedents and consequences of brand love. That is, neither academics nor practitioners have determined comprehensively what really drives brand love. In response to this gap, we propose a conceptual framework and hypotheses, based on an extensive literature review, in the next section, which we test with a mixed methods approach. First, we report on a qualitative study to confirm whether our proposed conceptual framework and hypotheses about the key drivers (e.g., functional, sensory, and communicative brand uniqueness) and behavioral consequences (e.g., brand loyalty, willingness to pay a price premium) of brand love align with real-world consumer feelings and experiences. Second, with a quantitative approach, we test the validity of the hypotheses. The results identify functional and sensory brand uniqueness as indirect antecedents of brand love; brand satisfaction, brand fit with the inner self, and personal experiences are direct antecedents. Contrary to our theoretically grounded expectations, communicative brand uniqueness and brand pleasure are not influential determinants of brand love in our model.

Literature review
The aim of our research is to establish a comprehensive model that comprises the most relevant antecedents and consequences of brand love. To identify these variables, we follow a two-step procedure. In an extensive literature review, we first search for the antecedents and consequences that are most often analyzed in the context of brand love. Then, we conduct a qualitative study to check whether our variable set is complete. In 14 semi-structured interviews, we surveyed consumers to understand their relations with their most loved brand in great detail. We find evidence for two additional variables (brand satisfaction and personal experiences), less studied in past research, which we include in our conceptual model of brand love.
To start our analysis, we reviewed research that refers to interpersonal love, object love, or brand love, which helped us develop a clear understanding of brand love as the core construct for this study. Then with a Web of Science search of Social Sciences Citations Index entries, published since 1956, we identified 60 articles with "brand love" in their titles. From this sample, we identified 51 publications that analyzed brand love and its causal relations to antecedents or consequences. Through a careful assessment of the variables included in these models, we found more than 150 differently named variables, and we considered their interrelations, overlaps, and connections to brand love. Through this multistep review of prior literature, we establish an initial foundation of variables that have been introduced as antecedents, constituent elements, or consequences of brand love.
The review also reveals that most brand love studies are narrowly focused, analyzing segments of a larger, causal network. In particular, some researchers concentrate on the antecedents of brand love (e.g., Huang 2019; Long-Tolbert and Gammoh 2012; Rauschnabel and Ahuvia 2014), whereas others examine its consequences (e.g., Albert et al. 2009;Rossiter 2012). Despite the many facets involved in the formation of brand love though, no prior study has provided a comprehensive framework detailing how it develops and exerts effects on brand performance. In Table 1, we provide an overview of the variables used in our study and show how they have been used in existing literature, sorted according to whether they appeared as antecedent variables, elements, or consequences of brand love. In the following sections we discuss the components of our conceptual model ( Fig. 1) in more detail.
Brand love as a core construct Fournier (1998) determines that consumers develop different bonds with the brands they use. At the extreme, strong consumer-brand relations are deep, lasting, and accompanied by intense feelings. Brand love represents the most intensive consumer-brand relationship (Rossiter 2012), distinct from other, more transient forms, such as flings or fads (Alvarez and Fournier 2012;Fournier 1998). Research into interpersonal love identifies deep affection (Sternberg 1986;Walster 1971) and separation distress (Hatfield and Sprecher 1986;Hazan and Shaver 1987;Rubin 1970) as fundamental components of the love construct. An established conceptualization likewise represents brand love as a two-dimensional construct, comprised of deep affection and separation distress (Bergkvist and Bech-Larsen 2010;Rossiter 2012). Affection distinguishes love from negative or neutral relationships, but it cannot differentiate it from mere liking, which also might evoke affection or positive feelings. Therefore, the second component is required, because separation distress, which arises when the loved object is absent, effectively differentiates love from pure liking. Sternberg (1986) accordingly proposes an "absence test" to indicate love, because a person's reaction to the absence of a beloved person distinguishes love from liking. According to Ahuvia (2015, p. 3), referring to consumers, "The more horrible they anticipate the loss to be, the more sure they are that they truly love that thing." We note though that separation distress also is not sufficient to assess brand love on its own, in that it might arise in relations marked by dependence, even if they do not feature love (Fournier 1998).
Beyond establishing its two constitutive components, we also delineate brand love by clearly distinguishing it from its antecedents and consequences. Bergkvist and Langner (2020) emphasize the serious threats that can arise for content validity, discriminant validity, and comparability if researchers measure a construct using items that capture its antecedents and consequences, rather than the target construct itself. Notably, antecedents of brand love often affect other constructs too, such as liking, attachment, or attitude. Furthermore, we recognize that outcome variables can reflect the impact of other constructs, not just the target construct. In contrast with these conceptual concerns though, brand love research often blurs these distinctions. For example,  include self-brand integration as a component of brand love, whereas other studies regard it as an important antecedent (Albert and Merunka 2013;Bergkvist and Bech-Larsen 2010;. More generally, a prevalent tendency is to include  Please see "Appendix 3" for details on all articles listed in Table 1 In this table we consider 46 studies. Single-construct factors and first/higher-order factors were assigned to a category on the basis of their (a) naming (naming identical or similar to the target variable) and (b) their operationalization (same or similar items as the target variable) brand love antecedents in measures of the brand love construct, instead of treating them as independent constructs (Albert et al. 2009;. Furthermore, some researchers rely on behavioral outcome variables (e.g., willingness to invest resources) to measure brand love, rather than treating them as consequences . Considering these potential issues and the lack of clarity in prior research, we adopt an operationalization of brand love provided by Bergkvist and Bech-Larsen (2010), Rossiter (2012), and, which explicitly seeks to minimize the overlap of brand love with related constructs and thereby ensures greater construct validity. It features both affection and separation distress as indicators, and it avoids the explicit use of the term "love" in a brand context, which might create confusion when consumers use that term loosely in consumption contexts ("I'd love to see that movie," "I love Cheerios," Rossiter 2012, p. 909).

Antecedents of brand love
Research on the antecedents of brand love tends to be restricted, focusing on just a few variables. For example,  examine high brand quality as the sole predictor of brand love (see also Hwang and Kandampully 2012).  investigate two constructs: hedonic and self-expressive benefits (see also Albert and Merunka 2013;Bergkvist and Bech-Larsen 2010). Rarely does research include brand-or product-specific features, related to brand design or brand communication, and their effects on brand love. In studies that consider product-related features, the variables tend to pertain to a more general level, such as product quality , or else the features (e.g., uniqueness, high price) are mentioned in items included in the brand love construct ). This tendency is surprising, considering that the "brand stimulus," which is composed of brand-related stimuli such as brand logo, packaging, product, or communications, is the primary origin of the cognitive, affective, and social processes within consumers (Fournier 1991). The importance of the brand stimulus is also confirmed by the results of our qualitative study, as summarized in two exemplary statements (Table 2): Simply because I think that they [MINI] make very good commercials. They really give the brand an outstanding character […] and they also have a certain sense of humor in their commercials. And they are cool. (Anna, female, 26).
On this note, the designs are also something extraordinary. Of course, they [Lyle and Scott] chose a unique logo which, in my opinion, no one can copy very easily. (Samuel,male,26).
Thus, in contrast with previous research, we consider the brand stimulus as the starting point of a causal chain that leads to brand love. In this chain, economic success requires brand uniqueness (Keller and Swaminathan 2019). Brands must establish unique attributes that differentiate them from competitors, to ensure that consumers prefer them. These so-called points of difference have their origin in either functional, performance-related or imagery-related brand perceptions (Keller 2003, p. 131). Virtually any functional, communicative, or sensory brand attribute can provide a foundation to establish brand uniqueness. Consequently, we include functional, sensory, and communicative brand uniqueness as the major exogenous variables in our model ( Fig. 1).
In terms of frequency, the most popularly studied brand love antecedents relate to consumers' identification with the brand (Table 1). Overall, 20 articles include variables such as "self-expressive brand"  or "self-brand image congruency" (Loureiro et al. 2017), with 13 studies classifying these variables as antecedents. The items that make up these variables usually measure the extent to which the brand fits with or reflects the consumer's personality. Ranking next in frequency are variables that measure pleasure linked to consumption, such as "hedonic value" or "pleasure," which appear in 11 studies (6 studies consider pleasure as an antecedent of brand love). Accordingly, our conceptual model includes brand fit with the consumer's inner self and brand pleasure. Porsche All qualitative interviews were conducted in German, and the author team translated the excerpts into English. Names of the respondents are changed, for anonymity Moreover, our qualitative study provides support for including two further antecedents: memories of meaningful past experiences the consumer has had with the brand, such as childhood, family, or travel experiences and satisfaction with the brand (Table 2). Prior research occasionally considers the impact of meaningful past personal experiences, and some evidence indicates its high relevance, in support of its inclusion in our framework . For example, Braun-LaTour et al. (2007, p. 51) illustrate the importance of personal experiences in the formation of brand love by quoting one of their study participants: "When I was about 4 years old, I remember going to the beach in my Dad's Toyota truck. The day stood out because it was my birthday, and my Dad took off work to bring me to the beach. I can smell and taste the saltwater breeze mixed with the exhaust of this off-road machine, and to this day, I have a fascination with Toyota trucks." Similarly, brand satisfaction has been shown to be a highly relevant antecedent of brand love ).

Consequences of brand love
We find greater consistency with regard to the variables used to measure the outcomes of brand love (Table 1). Studies typically assess consumers' intentions to engage in a particular behavior because of the brand love they feel. The most frequently cited consequences of brand love are brand loyalty (29 1 studies, with 25 studies classifying this variable as a consequence), positive word of mouth (WOM) (22 studies, with 21 studies classifying this variable as a consequence), willingness to pay a price premium, respectively, invest resources (8 studies, with 6 studies classifying this variable as a consequence), and willingness to forgive mistakes (7 studies, with all studies classifying this variable as a consequence). Due to their prominence, we include these variables in our conceptual model.

Brand love
Brand love is the core construct for our model. We maintain that it is not simply a stronger form of brand liking (Rubin 1973) but rather is a qualitatively different phenomenon, determined by the two causal-formative indicators (Bollen and Diamantopoulos 2017) of affection and separation distress (Bergkvist and Bech-Larsen 2010;Rossiter 2012). Formally,

H1
The deeper the affection consumers feel for a brand, the more they love the brand. H2 The more separation distress consumers anticipate to feel in the absence of a brand, the more they love the brand.

Antecedents of brand love
In contrast with previous research, we focus particularly on the brand stimulus and its impact on the cognitive, affective, and social processes involved in consumers' brand love formation. To specify the brand stimulus, we concentrate on the uniqueness of the brand (and its communication), which so far has only been included in a few brand love studies (though usually as an item in the brand love measure; Hegner et al. 2017). We consider three subdimensions of brand uniqueness-functional, sensory, and communicative-as distal antecedents, then conceive of brand satisfaction, brand pleasure, brand fit with inner self, and personal experiences as proximal antecedents of brand love (Fig. 1). The proximal antecedents affect brand love directly; the effects of the distal antecedents are at least partially mediated. Brand satisfaction, pleasure, and fit are endogenous, in the sense that brand management efforts can affect them directly, but past personal experiences are exogenous and cannot be determined directly by brand activities.
Because brand love is a formative construct, its antecedents should exert effects by influencing the two causal-formative indicators, affection and separation distress (Temme et al. 2014). To simplify our exposition, we do not develop separate hypotheses for the influences of the various antecedent variables on the two causal-formative indicators of brand love but instead assume comparable effects.

Brand uniqueness
Brand uniqueness is "the degree to which customers feel the brand is different from competing brands" (Netemeyer et al. 2004, p. 211). Uniqueness is essential for brand success and a core facet of brand equity (Aaker 2010); it originates from three main sources: functional brand characteristics, sensory brand features, and brand communication. For consumers, brand love is the result of strong functional Langner et al. 2015), sensory , and communicative (Pawle and Cooper 2006) characteristics. Therefore, we derive the following hypotheses about the direct effects of brand uniqueness on brand love: H3 The more consumers perceive the (a) functional, (b) sensory, and (c) communicative uniqueness of a brand, the more they love the brand.

Brand satisfaction
Brand satisfaction is an overall cognitive evaluation of the total brand experience (Delgado-Ballester and Munuera-Alemán 2001; Fullerton 2005). All three uniqueness factors likely affect brand satisfaction and thus inform brand love. Both  and Langner et al. (2015) emphasize the importance of high product quality as a prerequisite for brand love. Consumers also tend to justify their love for a brand by referring to its superior qualities, such as "exceptional performance" or the "good looking design" (Batra et al. 2012, pp. 3). Brand quality has been established as an important antecedent of consumer satisfaction too (Golder et al. 2012). Finally, Esch et al. (2006) find that brand image exerts a positive impact on brand satisfaction. Thus, unique brand associations, established by experiencing the brand as distinct in its sensory, communicative, or functional elements, might enhance consumer satisfaction. Considering that satisfaction results from a positive cognitive evaluation of a brand, which leads to positive emotions, it should affect brand love (Keh et al. 2007;Tsai 2013). This prediction further is in line with Roy et al.'s (2013) assertion that positive, cumulative consumer experiences of high satisfaction increase their emotional attachment to a brand over time. In summary, we expect the three brand uniqueness factors to influence brand satisfaction; brand satisfaction then may exert a positive influence on brand love. Thus, we predict that brand satisfaction mediates the impact of the brand uniqueness subdimensions on brand love: H4 The more consumers perceive (a) functional, (b) sensory, and (c) communicative brand uniqueness, the more they are satisfied with the brand. H5 The more consumers are satisfied with the brand, the more they love the brand. H6 Brand satisfaction mediates the effect of perceived (a) functional, (b) sensory, and (c) communicative uniqueness on brand love.

Brand pleasure
Brand pleasure refers to a summary judgment of how good it feels to interact with a brand (Le Bel and Dubé 1998). A consumer's evaluation of his or her hedonic experiences with the brand's sensory/emotive character is crucial Holbrook and Hirschman 1982). Moreover, superior functional aspects (e.g., good performance) can boost feelings of pleasure (Jordan 1998). Cho (2011, p. 140) reports a comment by a consumer that highlights this effect: "Apple makes my days easier. The unique system is different from the others and it makes me feel different too and proud of myself. The use of Apple was totally fun and nice." Consumers experience pleasure directly when they purchase or consume a brand and also indirectly by coming in contact with various brand communications (Brakus et al. 2009). Accordingly, previous research implies that all three brand uniqueness dimensions exert impacts on brand pleasure, which in turn is strongly linked to brand love Kim et al. 2008). In their qualitative study, Langner et al. (2015, p. 627) note that "many participants reported that they loved a brand because its usage evoked an extraordinarily good feeling." Similarly, Mugge et al. (2008) provide evidence that pleasure mediates the impact of superior product characteristics on product attachment. We expect that these findings about the mediating role of pleasure transfer to a brand love context too, so H7 The more consumers perceive (a) functional, (b) sensory, and (c) communicative uniqueness, the more they experience pleasure with the brand. H8 The more consumers experience pleasure with a brand, the more they love the brand. H9 Brand pleasure mediates the effect of consumers' perceived (a) functional, (b) sensory, and (c) communicative uniqueness on brand love.

Brand fit with inner self
Consumers often draw on the symbolic meaning of brands to form their self-identities, which helps them integrate into their social environments while also distinguishing themselves from others (Escalas and Bettman 2009). Such consumer-brand identification encompasses the "consumer's perceived state of oneness with a brand" (Stokburger-Sauer et al. 2012, p. 407). In the self-identification process, brand uniqueness is pivotal, in that it can fulfill consumers' need for uniqueness, defined as "an individual's pursuit of differentness relative to others that is achieved through the acquisition, utilization, and disposition of consumer goods for the purpose of developing and enhancing one's personal and social identity" (Tepper Tian et al. 2001, p. 52). Consumers prefer brands with unique properties that allow them to express their self-identities. Stokburger-Sauer et al. (2012) provide empirical evidence that brand uniqueness has a significant effect on brand identification.
Brand love literature also stresses the important role of brand identification in the formation of brand love   Kim et al. 2008;Ortiz and Harrison 2011;Tsai 2011). Among the origins of brand identification, it appears that personal brand fit is a key determinant of identifying with a brand. Therefore, consumers report that they fall in love with brands that express who they are (inner self) and who they want to be (desired self) Breazale and Ponder 2012). Reimann et al. (2012) also find that the integration of the brand with the self is significantly stronger in established brand love relationships than in recently formed ones. In neutral brand relationships-unlike brand love relationships-the levels of brand-consumer fit are consistently lower too. It appears that consumers first perceive initial similarities between their inner selves and a brand. Then over time, these perceptions of similarity strengthen, and brand integration progresses. Because the integration process has gone further in established brand love relationships, and the fit of the brand with the inner self is more prominent (cf. Breazale and Ponder 2012), we focus on this type of fit. All three proposed brand uniqueness features should strengthen fit with the inner self, which in turn may mediate the effects of brand uniqueness on brand love. Accordingly: H10 The more consumers perceive (a) functional, (b) sensory, and (c) communicative brand uniqueness, the more they perceive a fit of the brand with their inner self. H11 The more consumers perceive a fit of the brand with their inner self, the more they love the brand. H12 Consumers' perception of the fit of the brand with their inner self mediates the effect of perceived (a) functional, (b) sensory, and (c) communicative brand uniqueness on brand love.

Personal experiences
People can experience love for brands just because they connect the brand with meaningful moments or significant people from their personal lives (Albert et al. 2008;Breazale and Ponder 2012;Ortiz and Harrison 2011). In a study that asked consumers to describe critical incidents in their brand love formation process,  reveal that in addition to product experiences, participants cite emotioncharged experiences related to their loved brands. Yet these personal experiences were mostly separate from direct brand experiences. For example, more than half of the participants recalled specific childhood memories in which the brand played an important role. Brand-related events seem to occur at young ages in family environments and initialize brand socialization, which fosters brand familiarity and facilitates the development of close brand relationships. If brands are linked to specific people who have essential roles in consumers' lives, personal appreciation for these people transfers to the brands. For nostalgic consumers who long for things, persons, or situations from the past (Ortiz and Harrison 2011), brands also provide a means to maintain a sense of the personal past and revive valuable memories. Mugge et al. (2008) provide quantitative evidence of the positive effect of personal product memories on product attachment. Although only one quantitative study in the context of hotel branding ) to date has confirmed this impact of personal memories on brand love, the results from qualitative studies generally support such a relation . Therefore, we predict: H13 The more consumers relate personal experiences to a brand, the more they love the brand.

Consequences of brand love
To provide a comprehensive picture of brand love, we include brand loyalty, WOM intentions, willingness to pay a price premium, and willingness to forgive mistakes by the brand as consequences in our conceptual model (Fig. 1). Recent brand love research confirms generally positive effects of brand love on these outcomes. Moreover, the proximal antecedents of brand love (brand satisfaction, brand pleasure, and brand fit with inner self) may have positive, direct effects on these brand love consequences-in addition to the partially mediated effects by brand love. Therefore, we allow for both direct and indirect effects in our framework. However, for personal brand experiences, we anticipate that the effects on the consequences are fully mediated by brand love. We assume that meaningful experiences do not per se affect loyalty, willingness to pay a price premium, WOM intentions, or willingness to forgive mistakes, but they may exert impacts if they lead to brand love.
Brand loyalty is "a deeply held commitment to rebuy or repatronize a preferred product/service consistently in the future, thereby causing repetitive same-brand or same brandset purchasing, despite situational influences and marketing efforts having the potential to cause switching behavior" (Oliver 1999, p. 34). For brand managers, loyalty is an important behavioral outcome, because a loyal customer base can function as strong protection against competition and secure greater sales and revenues (Chaudhuri and Holbrook 2001;Delgado-Ballester and Munuera-Alemán 2001). Research shows that brand love enhances brand loyalty (Albert et al. 2009;Bergkvist and Bech-Larsen 2010;Tsai 2013). Moreover, satisfaction (Homburg et al. 2005), pleasure (Chaudhuri and Holbrook 2001), and identification (Stokburger-Sauer et al. 2012) generally have positive effects on brand loyalty.
Willingness to pay a price premium indicates the additional amount of money consumers would agree to pay for a product associated with a particular brand, rather than for a product without that brand link, when both brands offer comparable performance (Netemeyer et al. 2004). Such willingness is frequently reported as a result of brand love (Albert and Merunka 2013;Heinrich et al. 2012;Keh et al. 2007). Empirical findings indicate direct effects of satisfaction (Homburg et al. 2005), pleasure (Wakefield and Inman 2003), and identification (Homburg et al. 2009) on willingness to pay a price premium.
As another consequence of brand love, WOM intentions refer to "informal, person-to-person communication between a perceived non-commercial communicator and a receiver regarding a brand, product, an organization or a service" (Harrison-Walker 2001, p. 63). Positive WOM can facilitate brand success, as an influential communication channel (Keller 2007). It also is a significant consequence of brand love (Albert and Merunka 2013;. Rossiter (2012) finds that positive WOM happens approximately twice as often among consumers who love a brand than among those who merely like it. Again, satisfaction (Heitmann et al. 2007), pleasure (Ladhari 2007), and brand fit with inner self (Kuenzel and Halliday 2008) should have direct impacts on WOM intentions.
Like any actor, brands might make mistakes (Rusbult et al. 1991). Consumers might terminate brand relationships in response to brand misbehaviors, such as a product failures or ethical problems (Perrin-Martinenq 2004;Pullig et al. 2006), but they also might be willing to forgive some mistakes and to resist to negative information (Fournier 1998;Heinrich et al. 2012;). Consumers' willingness to forgive mistakes refers to a "willingness to give up retaliation, alienation, and other destructive behavior, and respond in constructive ways after an organizational violation of trust" (Xie and Peng 2009, p. 578). Consumers exhibit less sensitivity to brands' mistakes and accept mistakes more easily when they have an emotionally intensive bond with the brand (Donovan et al. 2012;Fedorikhin et al. 2008).
Thus, we anticipate 2 : H14 The more consumers love a brand, the more positive consequences toward the brand [(a) brand loyalty, (b) willingness to pay a price premium, (c) WOM intentions, (d) willingness to forgive mistakes] they show. H15 The more consumers (a) are satisfied with a brand, (b) experience pleasure with a brand, and (c) perceive a fit of the brand with their inner self, the more positive consequences [(a) brand loyalty, (b) willingness to pay a price premium, (c) WOM intentions, (d) willingness to forgive mistakes] toward the brand they show. H16 Brand love partially mediates the effects of (a) brand satisfaction, (b) brand pleasure, and (c) brand fit with inner self, and it fully mediates the effects of (d) personal experiences, on the consequences of brand love [(a) brand loyalty, (b) willingness to pay a price premium, (c) WOM intentions, (d) willingness to forgive mistakes].

Qualitative Study
The purpose of our qualitative study was to determine whether the identified antecedents and consequences of brand love and the hypothesized causal relationships are comprehensive and in line with real-world consumer feelings and experiences.

Sample and procedure
In total, 14 consumers (50% women), 26 to 57 years of age (M = 39 years), took part in semi-structured, face-to-face interviews (30-90 min in length), conducted in German. First, we provided participants with a short introduction to the concept of brand love. Second, we asked them to name and rank brands they love, according to the perceived strength of their love. Third, in the remainder of each interview, we focused on the brands the participants loved most.
The interviews included questions about their general relationships with brands, origins of brand love relationships, reasons they love brands, explorations of brand and product characteristics that lead to brand love, significant experiences with brands, roles others play in brand love relationships, behavioral outcomes of brand love relationships, and explorations of what would stop consumers from buying their most loved brands. Fourth, we analyzed participants' understanding of the brand love concept-that is, what brand love means to them. All of their responses and statements were provided in German; the author team translated them for use in this article.

Causal paths
We find qualitative evidence for the integrity of the causal paths of our conceptual model, across different participants and for different product categories. To present these findings, we use the pertinent example of one participant, Moritz (male, 27), who loves the notebook brand Sony Vaio due to its functional and sensory brand uniqueness (distal antecedents of brand love): Beautiful functionality-the [Sony Vaio] had beautiful, conspicuous, smart keys at the keyboard [and] was the first notebook of its generation that had a Blue-Ray Player.
These brand-endogenous antecedents of brand love affect pleasure, as a proximal brand love antecedent ("great pleasure that makes simply fun, happy"), which then elicits brand love through both affection ("a certain warmness") and separation distress: It broke by my own fault, and the moment I heard the terrible noises coming from it, I already missed it and felt very sad-not so much about the economic aspect but about losing a highly appreciated utility item, and I was instantly thinking about buying a new one.
Finally, this brand love positively affects Moritz's behavioral outcomes, such as his willingness to pay a price premium: Everyone knows that [the Sony Vaio] is expensive. You pay a certain amount for the brand and the image, although, you know, it could be two-thirds lower in price if there were a different label on it.…. Let's put it like this, I could have saved 600 Euro and bought another good product which in the end would have the same technical capabilities as the Sony Vaio but would not be so fancy, would not stand for the same thing.
Other participants described similar patterns, though the importance of particular paths varied, so for example, some participants stressed the importance of personal experiences, brand fit with inner self, or their brand satisfaction.

Completeness of variables
The respondents to this study frequently mentioned all the different variables we identified in our literature review.
Brand satisfaction and personal experiences represent two additional variables that have been studied occasionally before but that we also found in our qualitative study to be of substantial relevance for the concept of brand love. Therefore, we include them in our conceptual model. Table 2 provides some typical statements, reflecting the importance of the different constructs and their causal relations.
In particular, when talking about the loved brand, all respondents highlighted its uniqueness. In addition to high quality standards, superior functionality was an important discriminator of competing brands. Many participants emphasized the sensual properties of a loved brand and gave particular weight to distinct visual design.
However, with regard to communicative uniqueness, we observe some difference of opinions. That is, the majority of participants did not indicate that brand advertising communications were relevant to their brand love. However, some participants' statements indicated that loved brands can evoke clear, positive, unique associations, suggesting that communication is important to the development of personal feelings that induce brand love. In addition, five participants cited brand communication or highlighted the communicative uniqueness of their most loved brands. Because communicative uniqueness seems directly relevant to some consumers, we chose to retain this construct in our model and test it further with our subsequent quantitative analysis.
When describing brand love, the respondents mentioned highly emotional relations with their loved brand and even referred to it as an indispensable part of their life. Across the aggregated responses, we find that most respondents experience strong affection toward and separation distress with regard to their loved brand.
In reply to a question about why consumers might stop loving a brand, respondents mentioned both a failure in terms of fulfilling their expectations and disappointment. These comments imply that brand love is based on an exchange principle, whereby giving and taking are expected. In this sense, brand satisfaction is an important antecedent of brand love. We also identify frequent mentions of how the loved brand elicits feelings of brand pleasure, including joy, surprise, or happiness. Most respondents emphasize the importance of fit between the loved brand and their own identity, or else they describe the possibility of transferring desirable brand traits to their perception of themselves. Consumers use these loved brands to express themselves in their social environment. These findings from the qualitative study provide initial evidence that brand fit with inner self is an important driver of brand love. Finally, the participants often mentioned their personal experiences with the loved brand, including childhood or adolescent memories, travel, leisure activities, and gift situations, as well as experiences shared with family members or friends.
Turning to the consequences of brand love, respondents reported strong behavioral consequences as outcomes of their brand love, including expressions of brand loyalty and manifestations of their willingness to pay a price premium, talk positively about the brand, and forgive it for mistakes.

Quantitative study
With a second study, we test our proposed hypotheses regarding the antecedents and consequences of brand love quantitatively.

Sample
We surveyed 222 business students at a large university. Student samples are common in brand love research, for two reasons Gumparthi and Patra 2019;Sajtos et al. 2020). First, students are convenient to acquire and willing to participate simply in exchange for course credits. Second, as is true of interpersonal relationships, the quality of student-brand relationships should not differ from that of relations that other consumers develop with brands. Research in brand management also affirms that insights gained from student samples frequently transfer to other demographic groups Sattler 2006, 2007). After removing 23 questionnaires (due to non-native speakers, incomplete questionnaires, and one outlier) from the data set, our final sample comprises 199 participants, 108 of whom (54.3%) are women. The respondents have a mean age of 25 years (M = 24.94, SD = 3.34). We randomly assigned them to two conditions (brand love n = 132; brand liking n = 67), then merged the two groups into one data set to ensure sufficient variance in the responses.

Data collection
The questionnaire began with a short introduction, describing the purpose of the survey and the idea of brand love. After participants answered some demographic questions, we asked them to write down the brands they loved and to select the single brand they loved most. We gave participants in the brand liking group an additional instruction, namely, to list the brands they merely liked (and did not love) and to choose the brand they liked most from this set. We then asked participants in this latter group to ignore the brands they initially identified as loved brands for the remainder of the survey and evaluate only their most liked brand. We designed this procedure to ensure that the liked brands clearly contrasted with the loved brands selected during the first task. Next, participants in both groups answered questions related to key variables, in reference to their selected loved or liked brands, in the following sequence: hypothesized antecedents, hypothesized key construct of brand love, and hypothesized consequences. To avoid common method bias, we rotated three-item groups (Podsakoff et al. 2003). Finally, at the end of the questionnaire, we gave participants an opportunity to leave comments about the survey.

Measures
With the exception of brand love, we used reflective indicators for all constructs in our model. Affection and separation distress constitute brand love, so they function as causal-formative indicators of this focal construct. We used existing scales from literature wherever possible and adapted the item wording, as necessary, to reflect our research context. To specify the object of investigation clearly, each item used the term "my brand." We measured all indicators on 7-point Likert scales (1 = "completely disagree," 7 = "completely agree").
To translate the English-language measurement items obtained from previous research into German (see Appendixes 1 and 2 for the scale items in English and the reliability and validity statistics), we followed all the necessary steps suggested by Hambleton (2005). First, the items were translated into German by the first author. Second, a research assistant translated these German items back into English. Third, all authors checked whether the backtranslation matched the initial English items and resolved any differences through discussion. Fourth, the translated and refined questionnaire was pretested with nine students, who evaluated the comprehensibility of the items. If necessary, we adjusted the item formulations. The indicators of the three brand uniqueness variables (functional, sensory, communicative) were identical in their wording except for the particular uniqueness dimension captured by the item; we derived them from general uniqueness scales used in prior research (Albert et al. 2009;Netemeyer et al. 2004;Stokburger-Sauer et al. 2012) and assessed them with three items each. Because participants were free to select any brand they chose, there was great diversity in product categories. The uniqueness items (functional, sensory, communicative) logically should not apply equally to all these product categories. Therefore, in addition to using a 7-point Likert scale, we allowed participants to check a box marked "This statement does not apply to my brand." We adapted three items from prior literature to measure brand satisfaction (Brakus et al. 2009;Kuenzel and Halliday 2008;Oliver 1980). For brand pleasure, we considered five items, but we ultimately only used three items in the analysis, in line with our conceptual reasoning and previous operationalizations (Chaudhuri and Holbrook 2001;Mugge et al. 2006). In line with prior research, we also employed three items to measure brand fit with the inner self (Algesheimer et al. 2005;. Participants reported their personal experiences in connection with the brand on four items, one of which was negatively phrased (reverse polarity), which helps encourage participants' control and cognitive processing of the questionnaire items (it is not included in further analyses) (Podsakoff et al. 2003). We adapted three items, related to personal experience, from Albert et al. (2009) and Mugge et al. (2006). The measurement of brand love comprised the two causal-formative indicators of affection and separation distress, each of which we assessed with a single item, using wording from Bergkvist and Bech-Larsen (2010) and Rossiter (2012). We took three items to measure brand loyalty from existing scales: two items from Albert et al. (2009) and one item, already available in German, from Langner et al. (2009). We adopted an existing operationalization of two items to assess participants' willingness to pay a price premium (Netemeyer et al. 2004); we measured WOM intentions with three items from Price and Arnould (1999). To measure willingness to forgive the brand, we adopted four items from previous studies (Aaker et al. 2004;Heinrich et al. 2012). However, two items suggested by Aaker et al. (2004) focus on acceptance of weaknesses, so we ultimately did not include them in the analysis and instead used the two remaining items to measure willingness to forgive the brand.

Results
The results come from the complete data set, pooled across the two (brand love and brand liking) groups. Table 8 provides an overview of all tested hypotheses and the results concerning their acceptance and rejection.

Reliability and validity analysis
All measurement models were checked for reliability and validity (Table 3). In a first step, we computed Cronbach's alpha values for all reflective scales. They varied from 0.76 to 0.96, above the minimum reliability level of 0.70 (Nunnally 1978). In addition, the corrected item-to-total correlations ranged from 0.56 to 0.93, above the suggested threshold of 0.50 (Netemeyer et al. 2003). An exploratory factor analysis established unidimensionality for the indicators of each construct, with explained variances of 58.8-89.0% for the reflectively measured constructs. Subsequently, we subjected the items to a confirmatory factor analysis (CFA) using Mplus software (Muthén andMuthén 1998-2013). The CFA revealed a good overall fit. In addition, all factor loadings were highly significant (p < 0.001) and indicated a strong relationship between the items and their respective constructs (0.65-0.96). Indicator reliabilities (0.43-0.93) and composite reliabilities (0.76-0.96) provided satisfactory values, exceeding the minimum levels of 0.40 and 0.60, respectively (Bagozzi and Baumgartner 1994;Bagozzi and Yi 1988). Similarly, the average variances extracted (AVE) (0.56-0.89) exceeded the recommended threshold value of 0.50 (Fornell and Larcker 1981). Overall, the CFA results established convergent validity for all reflective scales. To assess discriminant validity, we applied the Fornell-Larcker criterion: For each pair of constructs, the squared correlation (see Table 3) was below the AVE for the constructs involved, thus demonstrating discriminant validity (Fornell and Larcker 1981). In conclusion, the reliability and validity analyses showed good psychometric properties for all measurement scales (for details, see "Appendix 2").

Model identification and fit
For our full structural equation model, we identified all reflective measurement models according to the three-indicator or two-indicator rule (Bollen 1989). We established the identification of the formative measurement model for brand love in two ways. First, we let brand love influence four reflectively measured behavioral outcome variables, which meets the 2+ emitted paths rule (Bollen and Davis 2009). Second, to establish a scale for the latent brand love variable, we specified a nonlinear constraint on the corresponding disturbance term, thereby standardizing brand love's variance to  unity (Edwards 2001). The proposed structural relationships between the constructs were recursive, so our entire model was identified. We produced maximum likelihood estimates for the parameters of the full structural model using Mplus software. We applied a bootstrap approach (5,000 bootstrap samples) to provide the standard errors and confidence intervals for all direct, indirect, and total effects and to take the non-normality of the data into account. The 90% confidence intervals (CI) corresponded to a one-tailed test at a 5% significance level. The overall fit statistics indicated that our model fit the data well: χ 2 (438) = 660.26, standardized rootmean-square residual = 0.06, comparative fit index = 0.95, Tucker-Lewis index = 0.94, and root-mean-square error of approximation = 0.05 (lower and upper bounds of the 90% CI were 0.042 and 0.058, respectively, and the p-value for the test of close fit was 0.45). Table 4 provides the standardized parameter estimates for the hypothesized direct effects with one-tailed p-values, along with the respective CI. Figure 2 depicts the estimated model with the standardized direct effects and corresponding significance levels.

Effects of brand love indicators on brand love overall
We find positive, highly significant effects of the causalformative indicators, affection (H1, γ = 0.43; p < 0.01) and separation distress (H2, γ = 0.44; p < 0.001), on brand love. The two effects exhibit a balanced influence on brand love, such that they are both substantial (Chin 1998). Together, the causal-formative indicators explain approximately 57% of the variance in the latent brand love variable.

Direct effects of distal antecedents on brand love indicators
After controlling for the indirect effects on the brand love indicators, through the more proximal causes of brand love (except the exogenous variable of personal experiences), we find no significant direct impact on affection (A) or separation distress (SD) due to the distal antecedents functional (H3a A , γ = 0.04, p = 0.316; H3a SD , γ = 0.12, p = 0.116), sensory (H3b A , γ = 0.13, p = 0.091; H3b SD , γ = 0.08, p = 0.196), or communicative (H3c A , γ = 0.07, p = 0.204; H3c SD , γ = − 0.07, p = 0.187) uniqueness. Thus, any significant effect of the uniqueness dimensions on brand love (or its behavioral consequences) appears completely mediated by the proximal antecedents, as we detail subsequently.

Direct effects of brand love and proximal antecedents on behavioral consequences
In line with our hypotheses, brand love exerts strong, highly significant effects on all four behavioral outcome variables. The standardized effect is greatest for the impact on brand loyalty (H14a, γ = 0.65, p < 0.001), closely followed by the effects on willingness to forgive mistakes (H14d, γ = 0.57, p < 0.001), willingness to pay a price premium (H14b, γ = 0.51, p < 0.001), and positive WOM intentions (H14c, γ = 0.42 p < 0.01). After controlling for the proximal antecedents' indirect effects through the brand love indicators, we identify strong, significant, direct effects of brand satisfaction on brand loyalty (H15aa, γ = 0.43, p < 0.001), willingness to pay a price premium (H15ab, γ = 0.29, p < 0.01), and WOM intentions (H15ac, γ = 0.29, p < 0.01). However, brand satisfaction does not directly influence willingness to forgive mistakes (H15ad, γ = − 0.17, p = 0.065). Brand pleasure relates directly to WOM intentions (H15bc, γ = 0.28, p < 0.01) but not to brand loyalty (H15ba, γ = − 0.02, p = 0.399), willingness to pay a price premium (H15bb, γ = 0.09, p = 0.235), or willingness to forgive mistakes (H15bd, γ = − 0.08, p = 0.251). According to our analysis, brand fit with inner self has no direct impact on any of the behavioral outcome variables: brand loyalty (H15ca, γ = − 0.00, p = 0.478), willingness to pay a price premium (H15cb, γ = 0.06, p = 0.227), WOM intentions (H15cc, γ = 0.01, p = 0.453), or willingness to forgive mistakes (H15cd, γ = 0.07, p = 0.253). Table 5 contains the standardized parameter estimates and bias-corrected bootstrap CI for the total effects in our model, which consist of direct and indirect effects. We start with reporting the total effects of the distal antecedents of brand love on its causal-formative indicators and on brand love overall which essentially are the result of the mediation processes via the proximal antecedents. Functional brand uniqueness has significant total effects on both affection (γ = 0.15) and separation distress (γ = 0.25) and also on brand love overall (γ = 0.17). Similarly, the total effects of sensory brand uniqueness on affection (γ = 0.17), separation distress (γ = 0.17), and brand love overall (γ = 0.15) are significant. In contrast, the total effects of communicative brand uniqueness on affection (γ = 0.12) and separation distress (γ = − 0.01) do not achieve significance. Thus, communicative brand uniqueness does not influence brand love overall (γ = 0.05). Most of the proximal antecedents of brand lovethat is, brand satisfaction (γ = 0.16), brand fit with inner self (γ = 0.17), and personal experiences (γ = 0.16)-show significant total effects on brand love overall. As implied by its lack of effects on brand love's causal-formative indicators though, the total overall effect of brand pleasure on brand love fails to reach significance (γ = 0.10).

Total and indirect effects of distal and proximal antecedents on brand love overall
Noting that the functional and sensory uniqueness of the brand exert positive overall effects on brand love despite non-significant direct effects, we consider these mediation processes more closely (Table 6). For functional brand uniqueness, the significantly positive total indirect effect on brand love appears to be caused through a mediation by brand satisfaction (γ = 0.06) and, to a slightly lesser extent, by brand fit with inner self (γ = 0.03). Sensory brand uniqueness significantly influences brand love through a mediation by brand satisfaction (γ = 0.03). In conclusion, brand satisfaction and brand fit with inner self fully mediate the overall effects of functional brand uniqueness (via affection and separation distress) on brand love (H6a: supported, H12a:

Total and indirect effects of distal and proximal antecedents on behavioral consequences
In addition to the total effects on brand love overall, we estimated the total effects of the distal and proximal antecedents of brand love on its behavioral consequences (Table 5). For the proximal antecedents (except for personal experiences), we also report their indirect effects on the behavioral consequences through the mediation of brand love (Table 7). The effect of the brand stimulus (i.e., distal antecedents of brand love) on the behavioral consequences is fully mediated, first by the proximal antecedents and then by brand love, thus, we report the total effects of the distal antecedents. The total effect (in this case, equal to the total indirect effect) of personal experiences on each behavioral outcome variable is significant (brand loyalty γ = 0.11; willingness to pay a price premium γ = 0.08; WOM intentions γ = 0.07; willingness to forgive mistakes γ = 0.09). This result corroborates the mediation hypotheses H16da-H16dd. Brand satisfaction exerts a significant total effect on brand loyalty (γ = 0.53), willingness to pay a price premium (γ = 0.37), and WOM intentions (γ = 0.35) but not on willingness to forgive mistakes (γ = − 0.08). The effects of brand satisfaction indirectly conveyed through brand love are significant for all behavioral consequences (including willingness to forgive mistakes), in support of H16aa-H16ad. With regard to brand pleasure, we find a significant total effect only for the relationship with WOM intentions (γ = 0.32); the total effects on the other three behavioral outcome variables are not significant (brand loyalty γ = 0.04; willingness to pay a price premium γ = 0.15; willingness to forgive mistakes γ = − 0.02). All indirect effects of brand pleasure on the behavioral consequences were not significant. Thus, we must reject H16ba-H16bd. Brand fit with inner self has a significant total effect on willingness to pay a price premium (γ = 0.15) and willingness to forgive mistakes (γ = 0.16) but not on brand loyalty (γ = 0.11) or WOM intentions (γ = 0.08). All the indirect effects of brand fit on behavioral consequences are significant, which supports the mediation hypotheses, H16ca-H16cd. We find significant total effects of both functional and sensory brand uniqueness on brand loyalty (γ = 0.27 and 0.16), willingness to pay a price premium (γ = 0.22 and 0.15), and WOM intentions (γ = 0.23 and 0.19). However, neither functional (γ = 0.03) nor sensory (γ = 0.04) brand uniqueness influences willingness to forgive mistakes. In line with the lack of effect on brand love, communicative brand uniqueness has no effect on the behavioral outcome variables (brand loyalty γ = 0.03; willingness to pay a price premium γ = 0.06; WOM intentions γ = 0.08; willingness to forgive mistakes γ = 0.02).
In conclusion, brand love overall (via affection and separation distress) mediates the effects of personal experiences on all four behavioral outcome variables (H16da-H16dd: Table 6 Total, direct and indirect effects of distal antecedents on brand love overall Standardized total, direct and indirect effects were estimated using Mplus 7; 5000 bootstrap samples have been used to provide biascorrected confidence intervals; Total effect = direct effect + total indirect effect; Direct effect = effect of distal antecedent on brand love overall via causal-formative indicators of brand love; (Total) indirect effect = Effect of distal antecedent on brand love overall via proximal antecedent(s) and causal-formative indicators of brand love supported). We also find that brand love overall partially mediates the significant effects of brand satisfaction on brand loyalty (H16aa: supported), willingness to pay a price premium (H16ab: supported), and WOM intentions (H16ac: supported). Whereas brand fit with inner self has no direct influence on the behavioral consequences, brand love fully mediates the effects of brand fit with inner self on willingness to pay a price premium (H16cb: supported) and willingness to forgive mistakes (H16cd: supported). The complete mediation of the effects of brand fit with inner self and personal experiences on willingness to forgive mistakes, in combination with the lack of other direct effects, indicates that brand love plays a dominant role in establishing consumers' willingness to forgive mistakes.
In line with the findings for the direct effects, brand pleasure is not as strong a predictor as we expected; all hypothesized mediation effects are not significant (H16ba-H16bd: rejected). The only direct effect of brand pleasure that we observe, independent of the brand love construct, is its effect on WOM intentions.

Discussion
Our comprehensive research model has been largely corroborated by the empirical results; it offers a valid description of the origin, nature, and impact of brand love. We thus establish a brand love construct, with its two causal-formative indicators of affection and separation distress. Whereas some researchers (Ahuvia et al. 2013) regard separation distress as one of many indicators of brand love, our findings confirm the conclusions offered by Bergkvist and Bech-Larsen (2010), Rossiter (2012), andLangner et al. (2015): Both affection and separation distress are significant, constitutive elements of brand love. Our model empirically confirms that the brand love construct is distinct from other, related constructs (e.g., satisfaction, pleasure, fit with inner self, loyalty). Relative to broader conceptualizations of brand love (e.g., , our study affirms the advantages of  Standardized total, direct and indirect effects were estimated using Mplus 7; 5,000 bootstrap samples have been used to provide bias-corrected confidence intervals; Total effect = direct effect + total indirect effect; Direct effect = effect of proximal antecedent of brand love on behavioral outcome; Total indirect effect = Effect of proximal antecedent of brand love on behavioral outcome via brand love overall Rejected a precise operationalization, using both affection and separation distress. The question of whether brand love is a formative or reflective measure is not trivial. 3 To gain further insights into the nature of brand love, we analyzed the relationship between affection and separation distress in more detail. These indicators correlate at r = 0.51, leading to a Cronbach's alpha of 0.67, below the common threshold of 0.70 (Nunnally 1978). Furthermore, in an unreported scatter plot we find no linear relation between the two brand love indicators. Instead, medium to high values of affection tend to be accompanied by high values of separation distress, whereas substantial variation in separation distress is observed at low values of affection. Thus, both rational arguments and our empirical findings suggest the formative instead of reflective nature of brand love. We confirm brand love's predictive value too, with notable links to four behavioral outcomes: loyalty, willingness to pay a price premium, WOM intentions, and willingness to forgive mistakes.
Among the antecedents, the results validate most of our hypotheses, but we are surprised that a few of them are not supported. Specifically, brand satisfaction, brand fit with inner self, and personal experiences with the brand are direct drivers of brand love, but we do not find that brand pleasure is a direct antecedent. Still, these findings are in line with existing research Langner et al. 2015) that shows that in addition to affective reasons such as emotional personal experiences, cognitive drivers such as satisfaction are important to the formation of brand love. Also contrary to our expectations, functional, sensory, and communicative uniqueness are not direct antecedents of brand love. Rather, functional uniqueness exerts its influence on brand love via brand satisfaction and brand fit with inner self, while sensory uniqueness conveys its impact solely via brand satisfaction. This evidence of complete mediation offers a strong indication that we included the most important mediators of the impact of the brand stimulus on brand love (Zhao et al. 2010). Regarding communicative uniqueness, it does not affect brand love or any of the behavioral consequences we include in our model. The findings of our qualitative study signal the ambiguity of brand communication as a predictor of brand love; only a few respondents mentioned brand communication as a source of their brand love. In rejecting our hypothesis that communicative uniqueness is an antecedent of brand love, we derive several insights. First, our findings are in line with existing empirical research ) that specifies functional and sensory brand uniqueness as essential drivers of brand love development. Second, brand communication may not be particularly important for brand love development, especially among consumers who love their brand due to their personal experiences. In such cases, brand associations likely develop primarily through meaningful personal experiences, independent of any company-issued communication. Third, even though the quantitative study did not find evidence for a causal relationship between communicative uniqueness and brand love, we believe that communication is nevertheless essential for creating brand awareness and image, which in turn are necessary for the establishment of any brand relationship.
We validate the brand love construct as the core mediator of the effects of the distal and proximal antecedents on behavioral outcomes. Whereas prior literature has narrowly examined this mediating role of brand love, we extend such findings to address the links of brand love with a broader array of antecedents and behavioral consequences. According to our results, brand love fully or at least partially mediates the significant effects of distal antecedents (functional brand uniqueness, sensory brand uniqueness) as well as proximal antecedents (brand satisfaction, brand fit to inner self, personal experiences) on behavioral outcomes of brand love (brand loyalty, willingness to pay a price premium, WOM intentions, willingness to forgive mistakes).
We find that after controlling for the influence of brand love, none of the antecedent variables exerts a significant direct impact on willingness to forgive mistakes. Thus, only brand love is able to produce the important consequence of willingness to forgive mistakes. This result provides support for the old saying that (brand) love is blind. Brand-loving consumers develop a tolerance for their brands' mistakes (Fournier 1998).
Although recent research reveals a direct effect of brand identification on loyalty (Homburg et al. 2009;Stokburger-Sauer et al. 2012) and an indirect effect of identification through brand love on loyalty (Bergkvist and Bech-Larsen 2010;Hwang and Kandampully 2012;Tsai 2011), we only partially observe similar effects for brand fit, as defined for this study. Despite an indirect effect of brand fit on brand loyalty, the total effects indicate that brand loyalty is influenced only by brand love and brand satisfaction, not by brand fit with inner self. Nor do we find any direct effects of brand fit with inner self on the behavioral consequences of brand love. Brand love, however, conveys a positive effect of brand fit with inner self on WOM intentions and willingness to forgive mistakes. That is, in our model, the effects of brand fit with inner self on these two behavioral consequences are fully mediated by brand love.
In summary, brand satisfaction, brand fit with inner self, and personal experiences are direct antecedents, and functional and sensory uniqueness are indirect antecedents, of the formation of brand love. In this context, brand satisfaction and brand fit with inner self are the key mediators in the causal chain from these direct and indirect antecedents to brand love. We confirm the strong mediating effect of brand love on all behavioral consequences. Even if communicative uniqueness and brand pleasure are not determinants of brand love, they are significant in relation to other constructs (e.g., effect of communicative uniqueness on brand fit with inner self; impact of brand pleasure on WOM intentions).

Conclusion, limitations, and implications
The purpose of this article has been to develop a holistic, causal model of brand love. Accordingly, our research contributes to a better understanding of the causal chain of brand love, from its formation to its behavioral outcomes. As other studies have before, our analysis confirms the strong impact that brand love exerts on behavioral consequences such as brand loyalty and willingness to pay a price premium. We find that functional and sensory brand uniqueness are indirect antecedents of brand love, whereas brand satisfaction, brand fit with the inner self, and personal experiences are direct antecedents. Contrary to our expectations, communicative uniqueness and brand pleasure do not influence brand love.
In terms of the limitations of this research, we acknowledge that the participants in our quantitative study are all students, though in the qualitative study, the respondents represent different age groups. Results obtained with student samples often transfer to other demographic groups (Lynch 1999;Sattler 2006, 2007), but a replication study with a population-representative sample could provide a better assessment of the potential influence of socio-demographic features. Furthermore, complexity and manageability considerations prevented us from integrating all possible antecedents and consequences into the proposed model. Continued research should investigate other potentially relevant variables (e.g., anthropomorphism, escapism, perceived corporate social responsibility, brand prestige), as well as examine the unexpected findings of our study, such as the lack of effects of brand pleasure or communicative brand uniqueness on brand love. A longitudinal research approach might help validate our model, in relation to the origin and development of brand love over time. The role of causality in structural equation modeling also can be controversial. For example, as a result of brand love, perceived brand uniqueness might increase, in which case uniqueness might be a consequence of brand love. Continued experiments thus should analyze the causality between uniqueness and brand love in more detail, to identify potential boundary conditions for the reverse causality.
Practitioners can apply our results in several ways. Overall, they should consider brand love as an important goal especially for the management of consumer goods, given its strongly positive effects on desirable behavioral outcomes: Consumers are more willing to pay higher prices, forgive the brand's mistakes, talk positively about the brand, and exhibit loyalty when they love a brand. To establish brand love relationships, we recommend four key steps.
First, managers should ensure their brands possess unique functional and sensory attributes. They are the cornerstones of brand love.
Second, customers need to be satisfied with the brand. In line with prior research, we emphasize the significance of satisfaction for brand love, and to enhance brand satisfaction, we again recommend that managers focus on functional and sensory uniqueness. Apple products provide a good example of this strategy, such that the iPhone's functional features consistently ensure high user friendliness, and its appealing aesthetics offer high sensory uniqueness that differentiate it from competitors' offerings.
Third, moving beyond brand satisfaction, which is not sufficient to establish brand love alone , managers should encourage consumers to identify with the brand. Brand fit with the inner self can be influenced through functional and communicative brand uniqueness, as when the MINI brand sought to become a lifestyle brand for young, urban consumers through its strong functional qualities (e.g., size, convenience) and marketing communications (e.g., portraying exciting lifestyles, aesthetics).
Fourth, a challenge for brand managers is finding a way to get consumers to establish meaningful personal memories that involve the brand, initiated by emotional events such as childhood, family, leisure, or vacation experiences. In this case, we recommend that brand managers communicate how their brands can provide a basis for such events. Such efforts could range from more brand-endogenous experiences, such as experiential flagship stores or sponsorships, to brand-exogenous experiences, such as including brands in childhood experiences (e.g., Mercedes bobby car), likely shared with their families (e.g., Cheerios), or enjoyed during special occasions (e.g., Tiffany & Co.).
Ultimately, even if brand love cannot be forced, insights into the development and management of brand love relationships can be used as strategic tools and inform more effective professional brand management. We encourage managers to use our findings and leverage our framework to manage consumers' love for their brands more effectively.